Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220218100910.35767-1-pbonzini@redhat.com>
In-Reply-To: <20220218100910.35767-1-pbonzini@redhat.com>
From:   Peter Gonda <pgonda@google.com>
Date:   Fri, 18 Feb 2022 12:44:07 -0700
Message-ID: <CAMkAt6rS-YrrSoYSU_9AukoTfrAy5awNFw3dkbCrFNoq0b3fWw@mail.gmail.com>
Subject: Re: [PATCH] selftests: KVM: add sev_migrate_tests on machines without SEV-ES
To:     Paolo Bonzini <pbonzini@redhat.com>
Cc:     LKML <linux-kernel@vger.kernel.org>, kvm list <kvm@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, Feb 18, 2022 at 3:09 AM Paolo Bonzini <pbonzini@redhat.com> wrote:
>
> I managed to get hold of a machine that has SEV but not SEV-ES, and
> sev_migrate_tests fails because sev_vm_create(true) returns ENOTTY.
> Fix this, and while at it also return KSFT_SKIP on machines that do
> not have SEV at all, instead of returning 0.
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Tested-by: Peter Gonda <pgonda@google.com>


>
> +#define X86_FEATURE_SEV (1 << 1)
> +#define X86_FEATURE_SEV_ES (1 << 3)

These conflict with these names but have different values:
https://elixir.bootlin.com/linux/latest/source/arch/x86/include/asm/cpufeatures.h#L402.
Is that normal in selftests or should we go with another name?
> +
>  int main(int argc, char *argv[])
>  {
> +       struct kvm_cpuid_entry2 *cpuid;
> +
> +       if (!kvm_check_cap(KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM) &&
> +           !kvm_check_cap(KVM_CAP_VM_COPY_ENC_CONTEXT_FROM)) {
> +               print_skip("Capabilities not available");
> +               exit(KSFT_SKIP);
> +       }
> +
> +       cpuid = kvm_get_supported_cpuid_entry(0x80000000);
> +       if (cpuid->eax < 0x8000001f) {
> +               print_skip("AMD memory encryption not available");
> +               exit(KSFT_SKIP);
> +       }
> +       cpuid = kvm_get_supported_cpuid_entry(0x8000001f);
> +       if (!(cpuid->eax & X86_FEATURE_SEV)) {
> +               print_skip("AMD SEV not available");
> +               exit(KSFT_SKIP);
> +       }
> +       have_sev_es = !!(cpuid->eax & X86_FEATURE_SEV_ES);
> +
>         if (kvm_check_cap(KVM_CAP_VM_MOVE_ENC_CONTEXT_FROM)) {
>                 test_sev_migrate_from(/* es= */ false);
> -               test_sev_migrate_from(/* es= */ true);
> +               if (have_sev_es)
> +                       test_sev_migrate_from(/* es= */ true);
>                 test_sev_migrate_locking();
>                 test_sev_migrate_parameters();
>                 if (kvm_check_cap(KVM_CAP_VM_COPY_ENC_CONTEXT_FROM))
> @@ -405,7 +440,8 @@ int main(int argc, char *argv[])
>         }
>         if (kvm_check_cap(KVM_CAP_VM_COPY_ENC_CONTEXT_FROM)) {
>                 test_sev_mirror(/* es= */ false);
> -               test_sev_mirror(/* es= */ true);
> +               if (have_sev_es)
> +                       test_sev_mirror(/* es= */ true);
>                 test_sev_mirror_parameters();
>         }
>         return 0;
> --
> 2.31.1
>